Chapter 01

Network Discovery & Asset Inventory

A device cannot be secured if it is not in the inventory. Discovery is the first substantive act in network security — and it's a measurement problem, not a scanner you press play on. Work through the lessons, practice on a simulated network where you can measure exactly how much you found, then run it for real.

① Lessons
② Learn (simulated)
③ Scan Mine (real)
The big idea. You never see the true device population D directly — only what each method reveals. The union of those is your discovered set D̂. What's left, D − D̂, is the discovery gap: invisible by definition. Every lesson below is one move in shrinking that gap honestly — and knowing what you still can't claim.
Simulated network — known ground truth

This network has a true population you can't see yet. Run the discovery methods in order; watch evidence merge into host records. Because the true count is known here (unlike a real network), you can measure exactly how complete and how sound your discovery was.

Discovery methods passive-first ordering
Live evidence stream
Merged host records
no hosts discovered yet
Discovery quality measurable here because ground truth is known
Completeness
—
found / real
Soundness
—
real / found
Discovery gap
—
1 − completeness
Step 1 — pick your system & copy the command
macOS
Windows
Linux
arp -a
Step 2 — paste the output
▸ what should it look like? (example)
🔒 everything runs in your browser — nothing is uploaded anywhere
Host records
IPMAC (masked)Vendor (inferred)Existence
Devices
0
0 filtered
Identified
0%
0/0
Randomized MAC
0
identity hidden
Your network map grouped by inferred device type
Ask AI about your network bring your own AI · nothing sent through us

Pick a question (or write your own). The tool packages your scan into a ready-to-use prompt — copy it into ChatGPT, Claude, or any AI assistant and get a plain-English answer. Your data goes only where you paste it.

analyze your network, then a copy-ready AI prompt appears here.
This builds the prompt locally from your parsed results — it doesn't call any AI itself, so it's free and private. (A built-in AI that answers in-app is a planned upgrade that needs a server.)
What this proves — and what it can't